Device maintenance method and device, computer device, and computer-readable storage medium

Abstract

The application relates to an equipment maintenance method and device, a computer device and a computer readable storage medium. The method comprises the following steps: performing prediction analysis on original electrical data of an electrical equipment at a plurality of historical moments to obtain target electrical data at a plurality of future moments after a current moment; performing fault analysis on the target electrical data at each future moment to predict a target fault probability of the electrical equipment at each future moment; determining a maintenance strategy of the electrical equipment at each future moment according to the target fault probability of each future moment; and performing maintenance management on the electrical equipment according to the maintenance strategy of each future moment. The method can reduce the probability of electrical equipment fault shutdown and improve the operation stability of a power system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power electronics, and particularly relates to a device maintenance method and device, computer equipment and computer readable storage medium. BACKGROUND

[0002] In a power system, reliable operation of electrical equipment is crucial to ensure the stability of power supply. With the continuous development of smart grid technology, real-time monitoring of operation data of electrical equipment has become an important means to improve the operation and maintenance efficiency of the power system.

[0003] The traditional operation and maintenance scheme of electrical equipment is to generate fault information of the electrical equipment when a fault of the electrical equipment is detected in real time, so that the electrical equipment is repaired by technical personnel. However, the stable operation of the power system will be affected when the electrical equipment is in fault shutdown. That is, the traditional operation and maintenance scheme of electrical equipment is difficult to adapt to the requirements of the power system for reliable operation of electrical equipment. SUMMARY

[0004] Therefore, it is necessary to provide a device maintenance method and device, computer equipment and computer readable storage medium capable of improving the operation reliability of electrical equipment in view of the above technical problems.

[0005] In a first aspect, the present application provides a device maintenance method, comprising:

[0006] predictively analyzing original electrical data of the electrical equipment at a plurality of historical moments to obtain target electrical data at a plurality of future moments after a current moment;

[0007] performing fault analysis on the target electrical data at each future moment to predict a target fault probability of the electrical equipment at each future moment;

[0008] determining a maintenance strategy of the electrical equipment at each future moment according to the target fault probability at each future moment;

[0009] performing maintenance management on the electrical equipment according to the maintenance strategy at each future moment.

[0010] In one of the embodiments, performing maintenance management on the electrical equipment according to the maintenance strategy at each future moment comprises:

[0011] when the maintenance strategy at each future moment comprises that the electrical equipment needs to be maintained at a target future moment and the maintenance manner is to reduce the operation power of the electrical equipment to a target operation power, reducing the operation power of the electrical equipment to the target operation power at the target future moment.

[0012] In one of the embodiments, the maintenance strategy of the electrical equipment at each future time is determined according to the target failure probability of each future time, including:

[0013] Obtaining the specified failure probability of at least one other electrical equipment outside the electrical equipment at each future time;

[0014] Determining the total operating power of the electrical equipment and the at least one other electrical equipment;

[0015] Determining the maintenance strategy of the electrical equipment at each future time according to the total operating power, the target failure probability of each future time and the specified failure probability of each future time.

[0016] In one of the embodiments, the original electrical data of the electrical equipment at a plurality of historical times is subjected to a prediction analysis to obtain target electrical data at a plurality of future times after the current time, including:

[0017] Performing feature analysis on the original electrical data of the plurality of historical times to obtain feature data of the plurality of historical times;

[0018] Correspondingly merging the original electrical data of the plurality of historical times and the feature data of the plurality of historical times to obtain related data of the plurality of historical times;

[0019] Classifying the related data of the plurality of historical times to obtain related data of the plurality of historical times under each category;

[0020] Determining the target electrical data of the plurality of historical times according to the related data of the plurality of historical times under each category;

[0021] Performing prediction analysis on the target electrical data of the plurality of historical times to obtain target electrical data of the plurality of future times.

[0022] In one of the embodiments, the target electrical data of the plurality of historical times is determined according to the related data of the plurality of historical times under each category, including:

[0023] Performing centralization processing on the related data of each historical time in each category of related data to obtain centralization data of each historical time of each category;

[0024] Determining the eigenvectors of the covariance matrix corresponding to the centralization data of each historical time of each category;

[0025] Determining the first preset number of values in the eigenvectors of each historical time of each category as the target electrical data of each historical time of each category.

[0026] In one of the embodiments, the target electrical data of the plurality of historical times is subjected to a prediction analysis to obtain target electrical data of the plurality of future times, including:

[0027] The target electrical data of the plurality of historical moments is used to train parameters in a preset prediction model, to obtain a trained target prediction model;

[0028] The target prediction model is used to process the target electrical data of the i th moment, to obtain the target electrical data of the i+1 th moment; wherein i is an integer greater than or equal to 1, and i+1 is less than or equal to the total number of moments of the plurality of future moments;

[0029] The target electrical data of the i+1 th moment obtained each time is determined as the target electrical data of each future moment.

[0030] In one of the embodiments, the original electrical data of the electrical equipment at the plurality of historical moments is subjected to prediction analysis, to obtain the target electrical data at the plurality of future moments after the current moment, including:

[0031] The random error and the p weight coefficients are obtained; wherein p is the number of moments of the plurality of historical moments;

[0032] The p weight coefficients are used to respectively perform weighted fusion on the target electrical data of the t th moment to the t+p-1 th moment, to obtain the t+p th target data; t is an integer greater than or equal to 1, and t+p is less than or equal to the total number of moments of the plurality of future moments;

[0033] The target electrical data of the t+p th moment is determined according to the t+p th target data and the random error;

[0034] The target electrical data of the t+p th moment obtained each time is determined as the target electrical data of each future moment.

[0035] In a second aspect, the present application provides a device maintenance device, the device comprising:

[0036] A prediction analysis module is configured to perform prediction analysis on the original electrical data of the electrical equipment at the plurality of historical moments, to obtain the target electrical data at the plurality of future moments after the current moment;

[0037] A fault analysis module is configured to perform fault analysis on the target electrical data of each future moment, to predict the target failure probability of the electrical equipment at each future moment;

[0038] A strategy determination module is configured to determine the maintenance strategy of the electrical equipment at each future moment according to the target failure probability at each future moment;

[0039] A maintenance management module is configured to perform maintenance management on the electrical equipment according to the maintenance strategy at each future moment.

[0040] In a third aspect, the present application provides a computer device, comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the method of any one of the above aspects when executing the computer program.

[0041] In a fourth aspect, the present application provides a computer readable storage medium, storing a computer program, and the computer program implementing the steps of the method of any one of the above aspects when executed by a processor.

[0042] In the embodiments of the present application, by predicting the target failure probability of the electrical equipment at each future time, determining the maintenance strategy of the electrical equipment at each future time according to the target failure probability of each future time, and performing maintenance management on the electrical equipment according to the maintenance strategy of each future time, the maintenance scheme for performing maintenance management on the electrical equipment can be determined in advance, and the electrical equipment is maintained according to the determined maintenance scheme, which helps to find the potential problems of the electrical equipment at the future time after the current time, effectively prevents the occurrence of failure, reduces the probability of electrical equipment failure downtime, and improves the operation stability of the power system. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other related drawings can be obtained by those skilled in the art without creative labor.

[0044] Figure 1 A structural schematic diagram of a device maintenance system provided for some embodiments;

[0045] Figure 2 A flowchart of a device maintenance method provided for some embodiments;

[0046] Figure 3 A flowchart of a method for determining the maintenance strategy of the electrical equipment at each future time provided for some embodiments;

[0047] Figure 4 A flowchart of a method for determining the target electrical data of multiple future times provided for some embodiments;

[0048] Figure 5 A flowchart of a method for determining the target electrical data of multiple future times provided for some other embodiments;

[0049] Figure 6 A structural schematic diagram of an intelligent monitoring system provided for some embodiments;

[0050] Figure 7 A flowchart of an acquisition method of original electrical data provided for some embodiments;

[0051] Figure 8 A flowchart of a method for determining whether an electrical device is faulty provided for some embodiments;

[0052] Figure 9 A flowchart of a method for determining whether an electrical device is faulty provided for some other embodiments;

[0053] Figure 10 A flowchart of a maintenance management method provided for some embodiments;

[0054] Figure 11 A structural diagram of a device maintenance apparatus provided for some embodiments;

[0055] Figure 12 A structural diagram of a computer device provided for some embodiments. DETAILED DESCRIPTION

[0056] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0057] In the description of the embodiments of the present application, the terms "first", "second" are only used for the purpose of description, and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0058] The device maintenance method provided by the embodiments of the present application can be applied to a device maintenance system. Figure 1 A structural diagram of a device maintenance system provided for some embodiments is as follows: Figure 1As shown, the equipment maintenance system 10 includes a plurality of electrical equipment 101 and a computer device 102 for obtaining original electrical data of the plurality of electrical equipment 101. The electrical equipment 101 communicates with the computer device 102, and the computer device 102 can display the target failure probability of each future time and / or the maintenance strategy of each future time. Taking a single electrical equipment 101 as an example, in some embodiments, the computer device 102 can directly collect the original electrical data of the electrical equipment 101; in other embodiments, the computer device 102 can obtain the original electrical data of the electrical equipment 101 through a sensor or other collection device; in still other embodiments, the original electrical data includes first type electrical data and second type electrical data, and the computer device 102 can directly collect the first type electrical data of the electrical equipment 101 and obtain the second type electrical data of the electrical equipment 101 through a sensor or other collection device.

[0059] The electrical equipment in the embodiments of the present application can include one of the following: a transformer, a circuit breaker, a motor, a capacitor, an Uninterruptible Power Supply (UPS) system, etc. Taking the electrical equipment as a transformer as an example, the original electrical data includes at least one of the following: winding temperature, oil temperature, load current, input voltage, output voltage, partial discharge information, vibration signal, etc. Taking the electrical equipment as a circuit breaker as an example, the original electrical data includes at least one of the following: operating current, switch state, voltage, current waveform, opening time, closing time, mechanical wear condition, temperature rise data, etc. Taking the electrical equipment as a motor as an example, the original electrical data includes at least one of the following: operating current, speed, temperature, vibration, power factor, torque, etc. Taking the electrical equipment as a generator as an example, the original electrical data includes at least one of the following: output current, output voltage, frequency, oil pressure, etc. Taking the electrical equipment as a capacitor as an example, the original electrical data includes at least one of the following: current, voltage, temperature, harmonic content, etc. Taking the electrical equipment as a power distribution cabinet as an example, the original electrical data includes at least one of the following: current, voltage, temperature, switch state, grounding resistance, leakage current, etc. Taking the electrical equipment as a UPS system as an example, the original electrical data includes at least one of the following: input voltage, output voltage, battery voltage, battery charge and discharge current, inverter temperature, frequency, etc.

[0060] Figure 2 A flowchart of a device maintenance method provided by some embodiments is shown in Figure 2 The method is applied to a computer device, and the method includes:

[0061] S201, performing prediction analysis on original electrical data of the electrical equipment at a plurality of historical time to obtain target electrical data at a plurality of future time after the current time.

[0062] The plurality of historical time points can be time points before the current time point or time points that have occurred before the current time point. In some embodiments, the time point can be replaced by one of the following: time step, time sequence, time period, time point, etc. In some embodiments, the original electrical data of a time point can be the electrical data collected at the time point, or the original electrical data of a time point can be the average, maximum, minimum, median, etc. of a plurality of electrical data collected in a time period between the time point and the previous time point of the time point.

[0063] In some embodiments, the interval between two adjacent time points in the plurality of historical time points can be the same as the interval between two adjacent time points in the plurality of future time points. In some embodiments, the interval between two adjacent time points in the plurality of historical time points can be the same as the interval between the last time point of the plurality of historical time points and the first time point of the plurality of future time points.

[0064] In some embodiments, the interval between two adjacent time points in the plurality of historical time points and / or the plurality of future time points can be in the range of 1 second to 24 hours, for example, the interval can be 1 second, 30 seconds, 1 minute, 1 hour, 12 hours, or 24 hours, etc.

[0065] In some embodiments, the number of time points in the plurality of historical time points can be the same as the number of time points in the plurality of future time points, or the number of time points in the plurality of historical time points can be greater than or less than the number of time points in the plurality of future time points.

[0066] In some embodiments, the parameter types in the target electrical data (e.g. parameter types such as current, voltage, etc.) can be the same as the parameter types in the original electrical data. In other embodiments, the parameter types in the target electrical data can be different from the parameter types in the original electrical data, for example, the parameter types in the target electrical data include the parameter types in the original electrical data and the parameter types corresponding to the feature data obtained by feature extraction on the original electrical data (e.g. parameter types such as current change, voltage change, etc.). For another example, the parameter types in the target electrical data include the parameter types corresponding to the feature data obtained by feature extraction on the original electrical data. For example, the feature data of each time point can be represented by a feature vector F = {f1, f2, …, f m} represents the values of different parameter types. m

[0067] In some embodiments, for an electrical parameter type (e.g. temperature) of an electrical device, in the case that there are a plurality of collected data (a plurality of temperature data) at a time point, the plurality of collected data can be fused to obtain a fusion result, and the fusion result of the time point is determined as the original electrical data of the time point.

[0068] ​Exemplarily, a weight corresponding to each of the plurality of collection data can be obtained, the plurality of weights are multiplied by the plurality of collection data respectively and then added to obtain a first result, the plurality of weights are added to obtain a second result, and a ratio of the first result to the second result is determined as the fusion result corresponding to the plurality of collection data.

[0069] For example, the following formula can be used The plurality of collection data are fused. Wherein, w i is a weight corresponding to each of the collection data, and n represents a total number of the plurality of collection data, x i represents each of the collection data.

[0070] S202, performing fault analysis on the target electrical data of each future time, and predicting a target fault probability of the electrical equipment at each future time.

[0071] The following describes some embodiments of S202:

[0072] In some embodiments, the target electrical data of each future time is input into a fault probability determination model, so that the fault probability determination model outputs the target fault probability of the electrical equipment at each future time.

[0073] In some embodiments, the following formula can be used The target electrical data of each future time is processed to obtain a calculation result of each future time, wherein Z represents the calculation result of each future time, X represents the target electrical data of each future time, μ represents a mean value of the target electrical data of each future time, and σ represents a standard deviation of the target electrical data of each future time. According to a comparison result of the calculation result of each future time and a first fault threshold, the target fault probability of the electrical equipment at each future time is determined. For example, in a case where the calculation result of a certain future time is greater than the first fault threshold, or in a case where the calculation results of a certain future time and a time after the future time are both greater than the first fault threshold, the target fault probability of the electrical equipment at the future time is determined to be 100%.

[0074] Exemplarily, for the target electrical data e i of each future time, when the number of times of the plurality of future times is n, the calculation formula of the mean value μ is The calculation formula of the standard deviation σ is

[0075] In some embodiments, the second failure threshold can be determined by μ+kσ, where μ represents the mean of the target electrical data at each future time, k represents a selected multiple, which can be 2 or 3, and σ represents the standard deviation of the target electrical data at each future time; and the target failure probability of the electrical equipment at each future time is determined according to the comparison result of the target electrical data at each future time and the second failure threshold. For example, in the case that the target electrical data at a future time is greater than the second failure threshold, or in the case that the target electrical data at a future time and the target electrical data at a time after the future time are both greater than the second failure threshold, the target failure probability of the electrical equipment at the future time is determined to be 100%. For example, the computer device obtains the power consumption of the electrical equipment at a future time to be 75 kW, and obtains the mean of the target electrical data at each future time to be 55 kW and the standard deviation to be 5 kW, then the second failure threshold determined by the selected multiple of 2 is 55+2x5=65 kW, since 75 kW is greater than 65 kW, the computer device determines that the electrical equipment fails at the future time, and the computer device can issue a warning to notify the maintenance personnel to check.

[0076] In some embodiments, an optimal hyperplane can be constructed according to the original electrical data at each historical time and the corresponding failure category to maximize the interval between categories, and the original electrical data at each historical time is divided into two categories of normal or failure by the optimal hyperplane; then, the target electrical data at each future time is analyzed for failure according to the optimal hyperplane to predict the target failure probability of the electrical equipment at each future time. In some embodiments, the computer device can determine the target failure probability of the electrical equipment at each future time according to the distance of the target electrical data at each future time relative to the optimal hyperplane. For example, the optimization target is the constraint condition is determining the optimal hyperplane to divide the original electrical data into two categories of normal or failure, where w is the weight vector of the hyperplane, b is the bias, and x i (i is 1 to n) is the sample data, y i (i is 1 to n) is the true label of the sample, represents all i.

[0077] In some embodiments, a matrix formed by the target electrical data of each future time can be subtracted by the reconstructed matrix (also referred to as the trained weight matrix) to obtain a target matrix; in a case where a square of a 2-norm of the target matrix corresponding to a certain future time is greater than a set threshold, it is determined that the electrical equipment fails at the future time. The reconstructed matrix is trained according to the original electrical data of a plurality of historical times. In some implementations, the computer device can determine a target failure probability of the electrical equipment at each future time according to a gap between the square of the 2-norm of the target matrix corresponding to each future time and the set threshold. For example, a calculation formula of a reconstruction error ReconstructionError is as follows: wherein the matrix X represents the target electrical data of each future time, is the reconstructed matrix, also referred to as the trained weight matrix, represents a 2-norm square of the matrix . The reconstructed matrix is trained according to the original electrical data of a plurality of historical times.

[0078] In some embodiments, in a case where the target electrical data at each future time includes a current value, a voltage value and a temperature value, the computer device can obtain a preset current range [I min ,I max ], a voltage range [V min ,V max ] and a temperature range [T min ,T max ], and in a case where the target electrical data at a certain future time satisfies a condition or or , it is determined that the electrical equipment fails at the time; wherein I, V and T respectively represent the current value, the voltage value and the temperature value at a certain future time. In some implementations, a first gap between the current value at each future time and the maximum current value I max , a second gap between the current value at each future time and the minimum current value I min , a third gap between the voltage value at each future time and the maximum voltage value V max , a fourth gap between the voltage value at each future time and the minimum voltage value V min , a fifth gap between the temperature value at each future time and the maximum temperature value T max , and a sixth gap between the temperature value at each future time and the minimum temperature value T min may be obtained, and a target failure probability of the electrical equipment at each future time is determined according to a smaller value of the first gap and the second gap, a smaller value of the third gap and the fourth gap, and a smaller value of the fifth gap and the sixth gap.

[0079] In some embodiments, a plurality of numerical ranges are determined, different numerical ranges in the plurality of numerical ranges correspond to different failure probabilities, for the jth future time, in a case that the target electrical data at the jth future time is in a target numerical range, a failure probability corresponding to the target numerical range is determined as a target failure probability of the jth future time. The target numerical range is a numerical range with the smallest range in the at least one numerical range, and the at least one numerical range is a numerical range in which the target electrical data at the jth future time is located.

[0080] In some embodiments, the computer device can further perform failure analysis on the target electrical data at each historical time to predict a target failure probability of the electrical device at each historical time. The manner of performing failure analysis on the target electrical data at each historical time is similar to the manner of performing failure analysis on the target electrical data at each future time, and details are not described herein.

[0081] S203, determining a maintenance strategy of the electrical device at each future time according to the target failure probability of each future time.

[0082] For example, the maintenance strategy includes 0, indicating that no maintenance is required. For another example, the maintenance strategy includes 1, indicating that maintenance is required and the target maintenance level is level 1. For another example, the maintenance strategy includes 2, indicating that maintenance is required and the target maintenance level is level 2.

[0083] In the plurality of maintenance levels, the higher the maintenance level, the smaller the probability of failure of the electrical device after maintenance of the electrical device using the maintenance level. For example, the current operating power of the electrical device can be reduced, and the higher the maintenance level, the more the current operating power of the electrical device is reduced. For example, when the target maintenance level is level 1, the current operating power of the electrical device is reduced to a first operating power, and when the target maintenance level is level 2, the current operating power of the electrical device is reduced to a second operating power, and the second operating power is less than the first operating power. For another example, the maintenance methods corresponding to the plurality of maintenance levels from low to high can include, in sequence, prompting inspection, reducing the operating power of the electrical device, cleaning, lubricating, stopping operation, and replacing the electrical device.

[0084] In some embodiments, S203 can include: in a case that the target failure probability of each future time is less than or equal to a probability threshold, determining that the maintenance strategy of the electrical device at each future time includes indication information that no maintenance is required.

[0085] In some embodiments, S203 can include: in a case where the target failure probabilities at the future time instants all have a value greater than the probability threshold, obtaining a maximum failure probability from the target failure probabilities at the future time instants, determining a probability difference value between the maximum failure probability and the probability threshold; obtaining a target future time instant at which the target failure probabilities at the future time instants first have a value greater than the probability threshold, determining a time difference value between the specific time instant corresponding to the maximum failure probability and the target future time instant; determining a target maintenance level for the target future time instant according to the probability difference value and the time difference value, and determining the maintenance strategies for the future time instants according to the target maintenance level for the target future time instant, so that the maintenance strategy for the target future time instant includes the indication information indicating that maintenance is needed and the target maintenance level, and the maintenance strategies for the other time instants among the future time instants except the target future time instant include the indication information indicating that maintenance is not needed. In some embodiments, the computer device can store a mapping relationship between the probability difference value, the time difference value, and the maintenance level, and the computer device can determine the target maintenance level according to the mapping relationship.

[0086] In some embodiments, S203 can include: obtaining a plurality of maintenance strategies preset in advance; updating the target failure probabilities of the electrical equipment at the future time instants to specific failure probabilities according to the plurality of maintenance strategies; and determining a target maintenance strategy from the plurality of maintenance strategies according to the specific failure probabilities at the future time instants, the maintenance cost corresponding to the target maintenance level, and the failure repair cost, so as to obtain the maintenance strategies for the future time instants.

[0087] In some embodiments, each maintenance strategy in the plurality of maintenance strategies corresponds to a future time instant at which maintenance is needed.

[0088] In some embodiments, different maintenance strategies represent maintenance of different maintenance levels at the same time instant among the future time instants, or maintenance of the same maintenance level at different time instants among the future time instants. For example, different maintenance strategies represent maintenance of a first maintenance level at different time instants among the future time instants.

[0089] For example, determining the target maintenance strategy from the plurality of maintenance strategies according to the specific failure probabilities at the future time instants, the maintenance cost corresponding to the target maintenance level, and the failure repair cost, includes: determining a plurality of cost values corresponding to the plurality of maintenance strategies according to the specific failure probabilities at the future time instants, the maintenance cost corresponding to the target maintenance level, and the failure repair cost, and determining the maintenance strategy corresponding to the minimum cost value in the plurality of cost values as the target maintenance strategy.

[0090] For example, the implementation manner of determining the cost value corresponding to each maintenance strategy can comprise: determining a total probability of the specific failure probability corresponding to each maintenance strategy at a plurality of future time points; multiplying the total probability by the failure repair cost to obtain a first cost; multiplying the level value corresponding to the target maintenance level by the failure repair cost to obtain a second cost; and adding the first cost and the second cost to obtain the cost value corresponding to each maintenance strategy.

[0091] For example, the determination of the maintenance strategy at each future time point can be performed by a linear programming method, wherein an optimization model of the linear programming method corresponds to a loss function wherein m i represents the maintenance decision at the i th time point (indicating the target maintenance level that needs to be maintained or does not need to be maintained, for example, m i = 0 indicates that the i th time point does not need to be maintained, or m i = 1 indicates that the maintenance level at the i th time point is 1, and m i = 2 indicates that the maintenance level at the i th time point is 2), f i represents the specific failure probability at each future time point determined according to the maintenance decision at each time point, for example, in the case where each maintenance decision indicates that no maintenance is needed, the specific failure probability at each future time point is the target failure probability, C m represents the maintenance cost corresponding to the target maintenance level indicated by the maintenance decision, C f represents the failure repair cost.

[0092] S204, according to the maintenance strategy at each future time point, performing maintenance management on the electrical equipment.

[0093] In some embodiments, S204 can comprise: in the case where the target failure probability at each future time point is less than or equal to the probability threshold value, outputting first prompt information according to the maintenance strategy at each future time point, the first prompt information being used to indicate that no maintenance is needed on the electrical equipment.

[0094] In some embodiments, S204 can comprise: in the case where the target failure probability at each future time point has at least one target failure probability greater than the probability threshold value, performing maintenance on the electrical equipment according to the maintenance strategy at each future time point.

[0095] In some embodiments, S204 can comprise: in the case where the target failure probability at each future time point has at least one target failure probability greater than the probability threshold value, outputting second prompt information according to the maintenance strategy at each future time point, the second prompt information being used to indicate that the electrical equipment is maintained at a target maintenance level at a target future time point. In this way, relevant personnel can perform maintenance on the electrical equipment at the target future time point according to the second prompt information.

[0096] In the embodiments of the present application, by predicting the target failure probability of the electrical equipment at each future moment, determining the maintenance strategy of the electrical equipment at each future moment according to the target failure probability at each future moment, and performing maintenance management on the electrical equipment according to the maintenance strategy at each future moment, the maintenance scheme for performing maintenance management on the electrical equipment can be determined in advance, and the electrical equipment is maintained according to the determined maintenance scheme, which helps to find potential problems of the electrical equipment at the future moment after the current moment, effectively prevents the occurrence of failure, reduces the probability of electrical equipment failure downtime, and improves the operation stability of the power system.

[0097] In some embodiments, performing maintenance management on the electrical equipment according to the maintenance strategy at each future moment can include: in the case that the maintenance strategy at each future moment includes that the electrical equipment needs to be maintained at the target future moment and the maintenance manner is to reduce the operating power of the electrical equipment to the target operating power, reducing the operating power of the electrical equipment to the target operating power at the target future moment.

[0098] Figure 3 A flowchart of a method for determining the maintenance strategy of the electrical equipment at each future moment is provided for some embodiments, as shown in Figure 3 The method is applied to a computer device, and the method includes:

[0099] S301, obtaining the specified failure probability of at least one other electrical equipment outside the electrical equipment at each future moment.

[0100] S302, determining the total operating power of the electrical equipment and the at least one other electrical equipment.

[0101] S303, determining the maintenance strategy of the electrical equipment at each future moment according to the total operating power, the target failure probability at each future moment, and the specified failure probability at each future moment.

[0102] In some embodiments of the present application, the sum of the operating powers of the electrical equipment and the at least one other electrical equipment should be greater than or equal to the total operating power.

[0103] Taking the electrical equipment as a first electrical equipment and the at least one other electrical equipment including a second electrical equipment as an example, if the failure probability of the first electrical equipment at the target future moment is 90%, the operating power of the first electrical equipment at the target future moment is 800W, the failure probability of the second electrical equipment at the target future moment is 0.5%, the operating power of the second electrical equipment at the target future moment is 800W, and the total operating power is 800W, the maintenance strategy of the first electrical equipment and the second electrical equipment at each future moment is to update the operating power of the first electrical equipment at the target future moment to 0, and keep the operating power of the second electrical equipment at the target future moment unchanged.

[0104] In some embodiments, the operation power of the electrical device and the at least one other electrical device can be scheduled by a linear programming method, wherein an optimization model of the linear programming method is a minimization objective value, the objective value is a result of multiplying and adding cost coefficients of the plurality of electrical devices (including the electrical device and the at least one other electrical device) and operation powers of the plurality of electrical devices; the linear programming method has a constraint condition including: the operation power of each device is greater than or equal to the minimum operation power of each device and less than or equal to the maximum operation power of each device; and the sum of the operation powers of the plurality of electrical devices is equal to the total operation power.

[0105] For example, the optimization model of the linear programming method corresponds to a loss function The corresponding constraint condition is ∑P i = P 最小总 . Wherein, P i is the power generation of the i-th device, c i is the cost coefficient of the i-th device. P i_min is the minimum operation power of the i-th device, P i_max is the maximum operation power of the i-th device, P 最小总 is the minimum total operation power.

[0106] Figure 4 A flowchart of a method for determining target electrical data of a plurality of future time points provided by some embodiments is shown in FIG. 1, which is applied to a computer device, and the method includes: Figure 4

[0107] S401, performing feature analysis on the original electrical data of the plurality of historical time points to obtain feature data of the plurality of historical time points.

[0108] ​In some embodiments, the feature data comprises at least one of: current fluctuation frequency data, voltage fluctuation frequency data, voltage jump rate data, temperature change rate data, frequency data of temperature being greater than a first temperature threshold, frequency data of temperature being less than a second temperature threshold, and the like. For example, if the computer device is able to obtain a plurality of current data from each historical time to the last historical time, the feature data of each historical time comprises current fluctuation frequency data. For another example, if the computer device is able to obtain a plurality of voltage data from each historical time to the last historical time, the feature data of each historical time comprises voltage fluctuation frequency data and / or voltage jump rate data. For another example, the original electrical data of each historical time comprises temperature data, and the feature data of each historical time comprises temperature change rate data. For another example, if the computer device is able to obtain a plurality of temperature data from each historical time to the last historical time, the feature data of each historical time comprises at least one of: frequency data of temperature being greater than a first temperature threshold, frequency data of temperature being less than a second temperature threshold, and the like.

[0109] S402, merge the original electrical data of a plurality of historical times and the feature data of a plurality of historical times correspondingly to obtain the related data of a plurality of historical times.

[0110] In some embodiments, after the feature data of each historical time is merged in the original electrical data of each historical time, the related data of each historical time is obtained.

[0111] S403, classify the related data of a plurality of historical times to obtain the related data of a plurality of historical times under each category.

[0112] The following description is one embodiment of classifying the related data of a plurality of historical times:

[0113] (1) initialize 2 center points or 3 center points. For example, 2 center points correspond to the following two categories: failure, normal. For example, 3 center points correspond to the following three categories: failure, warning, normal.

[0114] (2) calculate the distance between the related data of each historical time and each center point in the related data of a plurality of historical times, and assign the related data of each historical time to the nearest center point.

[0115] (3) update the position of each center point. Wherein, the new position of each center point is determined according to the following manner: add the related data of a plurality of historical times corresponding to each center point to obtain a third result; divide the third result by the vector length corresponding to the coordinates of each center point, and divide the third result by the vector length to obtain the coordinates of the new position of each center point. For example, the new position of each center point is wherein, ∑xi representing the center point c i corresponding to all historical moments i the sum, |c i | represents the vector length corresponding to the coordinates of each center point.

[0116] (4) Repeat steps (2) and (3) until the center point no longer changes or the maximum number of iterations is reached.

[0117] S404, according to the relevant data of each category under the plurality of historical moments, determine the target electrical data of the plurality of historical moments.

[0118] In some embodiments, S404 can include: centering the relevant data of each historical moment in each category of relevant data to obtain the centering data of each historical moment of each category; determining the eigenvectors of the covariance matrix corresponding to the centering data of each historical moment of each category; determining the first preset number of values in the eigenvectors of each historical moment of each category as the target electrical data of each historical moment of each category.

[0119] In some embodiments, centering the relevant data of each historical moment in each category of relevant data to obtain the centering data of each historical moment of each category can include: subtracting the average value of the relevant data of each historical moment in each category of relevant data from the relevant data of each historical moment in each category of relevant data to obtain the centering data of each historical moment of each category. For example, using X' = X - μ, the relevant data of each historical moment in each category of relevant data is centered to obtain the centering data of each historical moment of each category. Wherein, X represents the relevant data of each historical moment in each category of relevant data, μ represents the average value of the relevant data of each historical moment in each category of relevant data, and X' represents the centering data of each historical moment of each category.

[0120] In some embodiments, a predetermined matrix corresponding to the relevant data of each historical moment of each category can be obtained; the covariance matrix corresponding to the relevant data of each historical moment of each category is determined according to the product of the predetermined matrix and the transpose of the predetermined matrix; then, the eigenvalues and eigenvectors of the covariance matrix are calculated, thereby obtaining the eigenvectors of each historical moment of each category. For example, using determining the covariance matrix corresponding to the centering data of each historical moment of each category. Wherein, C represents the covariance matrix corresponding to the centering data of each historical moment of each category; A represents the predetermined matrix corresponding to the relevant data of each historical moment of each category; n represents the number of moments corresponding to the plurality of historical moments of each category. In this way, in the case of obtaining the covariance matrix, the eigenvalues and eigenvectors of the covariance matrix can be calculated.

[0121] S405, performing prediction analysis on the target electrical data of the plurality of historical moments to obtain target electrical data of a plurality of future moments.

[0122] In some embodiments, S405 can include: training parameters in a preset prediction model using the target electrical data of the plurality of historical moments to obtain a trained target prediction model; processing the target electrical data of the i th moment using the target prediction model to obtain the target electrical data of the i+1 th moment; wherein i is an integer greater than or equal to 1, and i+1 is less than or equal to the total number of moments of the plurality of future moments; determining the target electrical data of the i+1 th moment obtained each time as the target electrical data of each future moment. Wherein when i=1, the i th moment is the last moment of the plurality of historical moments.

[0123] For example, the target prediction model is used to process the target electrical data of the last moment of the plurality of historical moments to obtain the target electrical data of the 1 st future moment, then the target prediction model is used to process the target electrical data of the 1 st future moment to obtain the target electrical data of the 2 nd future moment, then the target prediction model is used to process the target electrical data of the 2 nd future moment to obtain the target electrical data of the 3 rd future moment, and so on. The target electrical data of the plurality of future moments can be obtained.

[0124] The following describes an embodiment of training parameters in a preset prediction model using the target electrical data of a plurality of historical moments to obtain a trained target prediction model:

[0125] The following formulas (1) to (6) correspond to the preset prediction model and the target electrical data x s of the s th historical moment

[0126] Forget gate f t : f s = σ(W f · [h s-1 , x s ] + b f ) (1).

[0127] Input gate i t : i s = σ(W i · [h s-1 , x s ] + b i ) (2).

[0128] Candidate value

[0129] Update cell state:

[0130] Output gate o t : o s = σ(W o · [h s-1 , x s ] + b o ) (5) ;

[0131] Update the hidden state:

[0132] In formulas (1) to (6), W f , W i , W C , W o represent weights, b f , b i , b C , b o represent biases, σ() is an activation function, and h() is a hyperbolic tangent function.

[0133] The weights and biases in the preset prediction model are trained using the following loss function: where MSE represents the mean square error loss function, represents the target electrical data at the s-th historical time predicted by the preset prediction model, n is the number of parameter categories included in the target electrical data, x s represents the real target electrical data at the s-th historical time. Exemplarily, the model parameters can be updated by a backpropagation algorithm to optimize the loss function.

[0134] Exemplarily, the historical current data of the electrical equipment is as follows: [10, 11, 12, 13, 14]. The trained target prediction model can predict the trend of the current at future time steps. For example, the predicted result is [15, 16, 17], which means that the current will gradually increase in the future operation of the equipment.

[0135] In this way, the device prediction unit can effectively determine the future trend of the electrical equipment and timely discover potential fault risks, thereby realizing more intelligent maintenance management and decision support.

[0136] Figure 5 A flowchart of a method for determining target electrical data at multiple future times is provided for another embodiment, as shown in Figure 5 The method is applied to a computer device, and the method comprises:

[0137] S501, obtain a random error and p weight coefficients; wherein p is the number of times of multiple historical times.

[0138] In some embodiments, the random error and the p weight coefficients can be pre-configured in the computer device. In other embodiments, the random error and the p weight coefficients can be trained according to the original electrical data of a plurality of early time points before a plurality of historical time points.

[0139] S502, using the p weight coefficients, respectively, to weight and fuse the target electrical data of the tth time point to the t+p-1th time point to obtain the t+pth target data.

[0140] Wherein, t is an integer greater than or equal to 1, and t+p is less than or equal to the total number of time points of the plurality of future time points.

[0141] Wherein, when t=1, the tth time point is the first time point of the plurality of historical time points, and the target electrical data of the tth time point to the t+p-1th time point is the original electrical data.

[0142] S503, determining the target electrical data of the t+pth time point according to the t+pth target data and the random error.

[0143] S504, determining the target electrical data of the t+pth time point obtained each time as the target electrical data of each future time point.

[0144] Using the method of the embodiments of the present application, using the p weight coefficients, respectively, to weight and fuse the target electrical data of the 1st time point to the pth time point to obtain the p+1th target data, and then according to the p+1th target data and the random error, determining the p+1th target electrical data (i.e. the target data of the 1st future time point), then using the p weight coefficients, respectively, to weight and fuse the target electrical data of the 2nd time point to the p+1th time point to obtain the p+2th target data, and then according to the p+2th target data and the random error, determining the p+2th target electrical data (i.e. the target data of the 2nd future time point), then using the p weight coefficients, respectively, to weight and fuse the target electrical data of the 3rd time point to the p+2th time point to obtain the p+3th target data, and then according to the p+3th target data and the random error, determining the p+3th target electrical data (i.e. the target data of the 3rd future time point), and so on, the target electrical data of the plurality of future time points can be obtained.

[0145] For example, the target electrical data of the t+pth time point can be determined according to the following formula: Wherein, α represents the bias, β1, β2 and βp represent the 1st weight coefficient, the 2nd weight coefficient and the pth weight coefficient (all included in the above p weight coefficients); Y p represents the 1st weight coefficient, the 2nd weight coefficient and the pth weight coefficient (all included in the above p weight coefficients); Y t , Y t+1 and Y t+p-1respectively represent the target electrical data at the tth (i.e., the tth) moment, the target electrical data at the t+1th (i.e., the t+1th) moment, and the target electrical data at the t+p-1th (i.e., the t+p-1th) moment; ò t represents a random error; represents the predicted target electrical data at the t+pth moment. Wherein, a, b1, b2, b p and ò t are trained by the original electrical data at a plurality of early moments before a plurality of historical moments.

[0146] For example, a, b1, b2, b loss function and the original electrical data at a plurality of early moments before a plurality of historical moments, a, b1, b2, b p and ò t are trained.

[0147] Wherein, represents the predicted electrical data at the rth early moment, Y r represents the original electrical data at the rth early moment, and N represents the number of parameter categories included in each original electrical data.

[0148] The electrical equipment intelligent monitoring system (i.e., the device maintenance system described above) based on artificial intelligence provided by the embodiments of the present application improves the operation safety and maintenance efficiency of electrical equipment through real-time data acquisition, intelligent analysis and fault prediction. Using artificial intelligence algorithms, the system can monitor the equipment operation state in real time, accurately predict future operation trends, and timely issue warnings to help effectively reduce equipment failure rate and improve equipment operation reliability and safety.

[0149] Figure 6 A structural schematic diagram of an intelligent monitoring system provided for some embodiments, as shown in Figure 6 The intelligent monitoring system 600 includes a data acquisition module 601, an operation monitoring module 602, an electrical data prediction module 603, a fault prediction module 604, a system management module 605, a remote monitoring module 606, and a central monitoring platform 607.

[0150] The data acquisition module 601 is configured to acquire the electrical equipment data information, and obtain original electrical data at a plurality of historical time points. The operation monitoring module 602 is configured to monitor the original electrical data of the electrical equipment at the plurality of historical time points. The electrical data prediction module 603 is configured to perform prediction analysis on the original electrical data of the electrical equipment at the plurality of historical time points, and obtain target electrical data at a plurality of future time points after a current time point. The fault prediction module 604 is configured to perform fault analysis on the target electrical data at each future time point, to predict a target fault probability of the electrical equipment at each future time point, and to perform fault analysis on the original electrical data at each historical time point, to predict a target fault probability of the electrical equipment at each historical time point. The system management module 605 is configured to determine a maintenance strategy of the electrical equipment at each future time point, to perform maintenance management on the electrical equipment according to the maintenance strategy at each future time point, and to determine a maintenance strategy of the electrical equipment at each historical time point, to perform maintenance management on the electrical equipment according to the maintenance strategy at each historical time point. The remote monitoring module 606 is configured to monitor the maintenance management of the electrical equipment. The central monitoring platform 607 is configured to obtain monitoring information of the remote monitoring module 606, and can also obtain the maintenance strategy at each future time point from the system management module 605, to perform maintenance management on the electrical equipment according to the maintenance strategy at each future time point.

[0151] Figure 7 A flowchart of a method for acquiring original electrical data is provided for some embodiments, as shown in Figure 7 , the method comprising:

[0152] S701, the data acquisition device acquires original electrical data of the electrical equipment.

[0153] In some embodiments, the data acquisition device can include a sensor.

[0154] S702, the data acquisition device transmits the original electrical data to the data storage.

[0155] S703, the computer device obtains the original electrical data from the data storage.

[0156] Figure 8 A flowchart of a method for determining whether the electrical equipment is faulty is provided for some embodiments, as shown in Figure 8 , the method is applied to a computer device, and the method comprises:

[0157] S801, original electrical data of the electrical equipment at a plurality of historical time points is obtained.

[0158] S802, target electrical data at the plurality of historical time points is determined according to the original electrical data at the plurality of historical time points.

[0159] S803, determine whether the electrical equipment fails at each historical moment according to the target electrical data of each historical moment.

[0160] S804, generate a failure analysis report in the case that the electrical equipment fails at a specified moment in the plurality of historical moments.

[0161] For example, in the case that the electrical equipment fails at the specified moment, the computer device can also output alarm information.

[0162] S805, in the case that the electrical equipment does not fail at each historical moment, continue to monitor the original electrical data.

[0163] By continuously monitoring the original electrical data, the original electrical data of the next batch of historical moments can be obtained.

[0164] Figure 9 A flowchart of a method for determining whether an electrical equipment fails is provided for some embodiments, as shown in Figure 9 The method is applied to a computer device, and the method comprises:

[0165] S901, obtaining original electrical data of the electrical equipment at a plurality of historical moments.

[0166] S902, performing prediction analysis on the original electrical data of the plurality of historical moments to obtain target electrical data at a plurality of future moments after the current moment.

[0167] S903, determining whether the electrical equipment fails at each future moment according to the target electrical data of each future moment.

[0168] S904, generating a failure analysis report in the case that the electrical equipment fails at a target future moment in the plurality of future moments.

[0169] For example, in the case that the electrical equipment fails at the target future moment, the computer device can also output alarm information.

[0170] S905, in the case that the electrical equipment does not fail at each future moment, continue to monitor the original electrical data.

[0171] Figure 10 A flowchart of a maintenance management method is provided for some embodiments, as shown in Figure 10 The method is applied to a computer device, and the method comprises:

[0172] S1001, obtaining target failure probabilities of each future moment.

[0173] S1002, determining a maintenance strategy of the electrical equipment at each future moment according to the target failure probabilities of each future moment.

[0174] S1003. Perform maintenance management on electrical equipment according to the maintenance strategy for each future time.

[0175] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, the steps corresponding to at least one optimization model in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the steps or stages of other steps.

[0176] The intelligent monitoring system for electrical equipment based on artificial intelligence provided in this application provides at least one of the following functions through the collaborative operation of multiple units:

[0177] Real-time monitoring: Key parameters of electrical equipment (such as current, voltage, and temperature) are collected in real time through the data acquisition unit. Utilizing artificial intelligence algorithms and big data analysis in the equipment operation monitoring unit, the operating status and abnormal conditions of the equipment can be detected in real time. Real-time data monitoring can issue timely warnings when abnormalities occur in electrical equipment, preventing potential faults from escalating into serious safety accidents.

[0178] Automatic alarm and fault analysis: When the system detects equipment abnormalities, it will automatically generate an abnormality report through the equipment operation monitoring unit and respond quickly through the abnormality alarm mechanism, effectively reducing the losses caused by equipment failures and ensuring the safety of the electrical system.

[0179] Equipment Failure Prediction: Utilizing an equipment failure prediction unit and combining it with a data trend prediction model, the system can predict future state changes of electrical equipment over time based on historical data and current status. Through deep learning and time series prediction models, the system can identify potential failure points and take preventative maintenance measures.

[0180] Preventative maintenance: Through prediction and analysis, the system can not only issue early warnings when equipment is about to fail, but also generate maintenance plans to proactively schedule preventative maintenance. This data-driven maintenance model is more scientific than traditional scheduled maintenance, effectively extending equipment lifespan and reducing economic losses caused by downtime due to malfunctions.

[0181] Data fusion and optimized scheduling: The system management unit integrates device operation monitoring and fault prediction results, and optimizes the operation scheduling of the device through data analysis. Using intelligent algorithms and data fusion technology, the system can effectively allocate power resources in power grid load management, reduce energy waste, and ensure that electrical equipment works in the best operating state.

[0182] Remote monitoring and intelligent management: The remote monitoring unit enables the system to realize centralized monitoring and management of distributed electrical equipment. By remotely obtaining device operation state data, management personnel can real-time master the operation state of the device without going to the site, conduct intelligent scheduling and optimization configuration, and improve the efficiency and accuracy of operation and maintenance.

[0183] All-round monitoring and analysis: Through monitoring, fault analysis, prediction and maintenance management covering the whole life cycle of the device, the system can optimize the whole life cycle management strategy of the device. From installation, operation, maintenance to device scrapping, the system can always provide all-round data support, effectively improve the operation efficiency of the device, prolong the service life of the device, and reduce the operation and maintenance cost of the device in the whole life cycle.

[0184] The embodiment of the application provides an intelligent monitoring system for electrical equipment based on artificial intelligence, which can realize real-time monitoring, fault analysis, fault prediction, remote management and maintenance scheduling of electrical equipment. The system mainly comprises a data acquisition unit, a device operation monitoring unit, a device fault prediction unit, a system management unit and a remote monitoring unit. The following describes some implementation modes of these units.

[0185] Implementation of the data acquisition unit: The data acquisition unit is used to acquire the operation state data of the electrical equipment, including but not limited to at least one parameter of current, voltage, temperature, load rate, power factor and the like. The unit acquires real-time operation data of the electrical equipment through sensors and data collectors, and transmits the data to a central processing module for analysis. The exemplary implementation steps can include: installing various sensors on the electrical equipment for monitoring key operating parameters, including voltage sensors, current sensors, temperature sensors and the like. The data acquisition unit transmits real-time data to the central processing system through the collector, which can be connected to the central processing module in a wired or wireless manner to ensure the real-time nature of data acquisition. All collected data will be stored in a data storage library as basic data for subsequent fault analysis, prediction analysis and remote monitoring.

[0186] Implementation of the device operation monitoring unit: The device operation monitoring unit utilizes artificial intelligence algorithms and big data technology to analyze the collected device operation data in real-time, identifying the normal operation state and potential failure risks of electrical equipment. Example implementation steps can include: First, establish a data model for electrical equipment by combining real-time data collected by the data acquisition unit with historical data. The model is self-adapted according to different types of electrical equipment to ensure the accuracy of the analysis. Use big data analysis technology to classify and process real-time data, extracting feature values of the device operation state. Analyze the health status of the device, including normal state, slight abnormality, and severe abnormality, through feature extraction algorithms. Use artificial intelligence algorithms, especially neural network-based classification algorithms, to further analyze the abnormal state of the device. The algorithm can learn from historical failure data to identify complex device anomalies. When an abnormal state is detected, the operation monitoring unit generates a failure analysis report showing the type of device anomaly, possible causes, and recommended maintenance solutions.

[0187] Implementation of the device failure prediction unit: The device failure prediction unit is based on the analysis results of the device operation monitoring unit, combined with data trend prediction technology, using a deep learning model to predict the future operation state of electrical equipment. Example implementation steps can include: First, collect historical data and real-time operation data of electrical equipment, input them into the deep learning model, and use these data to train and optimize the model. Use a deep learning model suitable for time series data to analyze the trend of the state of electrical equipment over time. The model predicts the future trend of key parameters (such as voltage, current, temperature, etc.) of electrical equipment to determine the future state of the device. If the predicted data exceeds the normal operating parameter range of the device, the system will generate a failure prediction report and issue an alarm in advance to prevent possible future failures of the device.

[0188] Implementation of the system management unit: The system management unit is responsible for managing the results of operation monitoring, failure analysis, and failure prediction, and optimizing the maintenance and scheduling of devices. Example implementation steps can include: The management unit first generates an intelligent maintenance management plan based on the device operation state, failure analysis report, and failure prediction results. The plan includes regular maintenance time, maintenance steps, and possible replacement parts. Based on data analysis results, the management unit can adjust the scheduling of electrical equipment to optimize the operating efficiency of the device. The system automatically generates scheduling data for adjusting load distribution in the power system. Device operation data and maintenance scheduling data are fused to form an electrical equipment database. This database is used to continuously update the device operation model, improving the self-learning and optimization capabilities of the system.

[0189] Implementation of the remote monitoring unit: the remote monitoring unit can obtain the running state data of the electrical equipment in real time, monitor the power consumption of the equipment, and remotely schedule through the data fusion result. The exemplary implementation steps can include: the remote monitoring unit transmits the power consumption data and the running state data of the electrical equipment to the central monitoring platform in real time through wireless communication technology or Internet connection. The central monitoring platform generates the running report and the power consumption state monitoring result of the equipment by fusing the historical data, real-time data and prediction data of the equipment. When the running state of the electrical equipment is abnormal, the system can issue an alarm through the remote platform and automatically generate a fault analysis report. For minor faults, the remote monitoring system can trigger an automatic repair program or remotely schedule tasks.

[0190] Based on the same inventive concept, the embodiments of the present application also provide a device maintenance apparatus for implementing the device maintenance method described above. The implementation scheme of the problem solving provided by the apparatus is similar to the implementation scheme described in the above method, so the specific limitations in one or more device maintenance apparatus embodiments provided below can refer to the limitations of the device maintenance method in the above, which will not be described here again.

[0191] In one exemplary embodiment, Figure 11 The structural schematic diagram of the device maintenance apparatus provided for some embodiments is shown in Figure 11 As shown in the figure, the device maintenance apparatus 1100 includes:

[0192] The prediction analysis module 1101 is configured to perform prediction analysis on the original electrical data of the electrical equipment at a plurality of historical time points to obtain target electrical data at a plurality of future time points after the current time point.

[0193] The fault analysis module 1102 is configured to perform fault analysis on the target electrical data at each future time point to predict the target fault probability of the electrical equipment at each future time point.

[0194] The strategy determination module 1103 is configured to determine the maintenance strategy of the electrical equipment at each future time point according to the target fault probability of each future time point.

[0195] The maintenance management module 1104 is configured to perform maintenance management on the electrical equipment according to the maintenance strategy of each future time point.

[0196] In some embodiments, the maintenance management module 1104 is further configured to, in the case that the maintenance strategy of each future time point includes that the electrical equipment needs to be maintained at a target future time point and the maintenance manner is to reduce the running power of the electrical equipment to a target running power, reduce the running power of the electrical equipment to the target running power at the target future time point.

[0197] In some embodiments, the policy determination module 1103 comprises: a failure probability acquisition unit and a maintenance policy determination unit; the failure probability acquisition unit is configured to acquire a specified failure probability of at least one other electrical equipment outside the electrical equipment at each future time; and the maintenance policy determination unit is configured to determine a total operating power of the electrical equipment and the at least one other electrical equipment, and to determine a maintenance policy of the electrical equipment at each future time according to the total operating power, a target failure probability at each future time, and the specified failure probability at each future time.

[0198] In some embodiments, the prediction analysis module 1101 comprises: an analysis unit, a merging unit, a classification unit, an electrical data determination unit, and a prediction unit; the analysis unit is configured to perform feature analysis on the original electrical data at the plurality of historical times to obtain feature data at the plurality of historical times; the merging unit is configured to merge the original electrical data at the plurality of historical times and the feature data at the plurality of historical times correspondingly to obtain related data at the plurality of historical times; the classification unit is configured to classify the related data at the plurality of historical times to obtain the related data at the plurality of historical times under each category; the electrical data determination unit is configured to determine target electrical data at the plurality of historical times according to the related data at the plurality of historical times under each category; and the prediction unit is configured to perform prediction analysis on the target electrical data at the plurality of historical times to obtain target electrical data at the plurality of future times.

[0199] In some embodiments, the electrical data determination unit comprises: a centralization processing subunit, a feature vector determination subunit, and an electrical data determination subunit; the centralization processing subunit is configured to perform centralization processing on the related data at each historical time in each category of related data to obtain centralization data at each historical time of each category; the feature vector determination subunit is configured to determine a feature vector of a covariance matrix corresponding to the centralization data at each historical time of each category; and the electrical data determination subunit is configured to determine a first preset number of values in the feature vector at each historical time of each category as the target electrical data at each historical time of each category.

[0200] In some embodiments, the prediction unit comprises a training subunit and an electrical data processing subunit; the training subunit is configured to train parameters in a preset prediction model using the target electrical data at the plurality of historical times to obtain a trained target prediction model; the electrical data processing subunit is configured to process the target electrical data at the i th time using the target prediction model to obtain the target electrical data at the i+1 th time; i is an integer greater than or equal to 1, and i+1 is less than or equal to the total number of times of the plurality of future times; and the electrical data processing subunit is further configured to determine the target electrical data at each future time as the target electrical data at the i+1 th time obtained each time.

[0201] In some embodiments, the prediction analysis module comprises a weight coefficient acquisition unit, a fusion unit, and an electrical data determination unit; the weight coefficient acquisition unit is configured to acquire random errors and p weight coefficients; p is the number of historical time points; the fusion unit is configured to use the p weight coefficients to perform weighted fusion on target electrical data at the tth time point to the (t+p-1)th time point, to obtain target data at the (t+p)th time point; t is an integer greater than or equal to 1, and t+p is less than or equal to the total number of future time points; the electrical data determination unit is configured to determine target electrical data at the (t+p)th time point according to the target data at the (t+p)th time point and the random errors; and the electrical data determination unit is further configured to determine the target electrical data at the (t+p)th time point obtained each time as target electrical data at each future time point.

[0202] Each module in the device maintenance apparatus described above can be implemented wholly or partially by software, hardware, and combinations thereof. Each module described above can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in a computer device in software form, so as to be called and executed by a processor to perform operations corresponding to each module.

[0203] In one exemplary embodiment, a computer device is provided, Figure 12 A structural schematic diagram of a computer device provided for some embodiments includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, the memory, and the input / output interface are connected through a system bus, and the communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to perform wired or wireless communication with external terminals, and the wireless communication can be achieved through WIFI, mobile cellular network, near field communication (NFC), or other technologies. The computer program is executed by the processor to implement a device maintenance method. The display unit of the computer device is configured to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer overlaid on the display screen, or can be a key, trackball, or touchpad arranged on the shell of the computer device, or can be an external keyboard, touchpad, or mouse, etc.

[0204] Those skilled in the art can understand that Figure 12 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0205] For example, in an exemplary embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program: performing prediction analysis on original electrical data of an electrical device at a plurality of historical time points to obtain target electrical data at a plurality of future time points after a current time point; performing fault analysis on the target electrical data at each future time point to predict a target failure probability of the electrical device at each future time point; determining a maintenance strategy of the electrical device at each future time point according to the target failure probability at each future time point; and performing maintenance management on the electrical device according to the maintenance strategy at each future time point.

[0206] In an embodiment, a computer readable storage medium is provided, and a computer program is stored in the computer readable storage medium, and the computer program is executed by a processor to implement the steps of the method provided in any of the above embodiments.

[0207] For example, in an exemplary embodiment, a computer readable storage medium is provided, and a computer program is stored in the computer readable storage medium, and the computer program is executed by a processor to implement the steps of the method provided in any of the above embodiments.

[0208] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments of the method.

[0209] The processor, each functional module or each functional unit in any of the embodiments of the present application can include an integration of any one or more of the following: a general purpose processor, an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a neural-network processing units (NPU), a controller, a microcontroller, a microprocessor, a programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, a data processing logic device based on quantum computing, an Artificial Intelligence (AI) processor, and the like. The general purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0210] The memory or computer readable storage medium in any of the embodiments of the present application can include at least one of a non-volatile memory and a volatile memory. The non-volatile memory includes integration of one or more of the following: Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Ferromagnetic Random Access Memory (FRAM), Flash Memory, magnetic random access memory, optical disk, Compact Disc Read-Only Memory (CD-ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, volatile memory, and the like. The volatile memory includes integration of one or more of the following: Random Access Memory (RAM) or external cache memory, and the like. As an illustration but not limitation, the RAM can be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), and the like.

[0211] Any combination of the technical features in the above embodiments can be made, and for the sake of brevity, not all possible combinations of the technical features in the above embodiments are described, however, as long as the combination of the technical features does not exist contradictions, it should be considered as the scope of the present application.

[0212] The above-described embodiments are merely illustrative of several embodiments of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

Claims

1. A device maintenance method characterized by, The method comprises: corresponding merging of original electrical data of a plurality of historical time points and feature data of the plurality of historical time points to obtain relevant data of the plurality of historical time points; initializing three center points, the three center points corresponding to the following faults, warnings and normal respectively; calculating the distance between the relevant data of each historical time point and each center point, and assigning the relevant data of each historical time point to the nearest center point; corresponding addition of the relevant data of a plurality of historical time points corresponding to each center point to obtain a third result; dividing the third result by the vector length corresponding to the coordinates of each center point to obtain the coordinates of the new position of each center point; continuing to perform the calculation of the distance between the relevant data of each historical time point and each center point until the center points no longer change or the maximum number of iterations is reached; subtracting the average value of the relevant data of a plurality of historical time points corresponding to each category from the relevant data of each historical time point in each category to obtain the centralized data of each historical time point in each category; determining the eigenvectors of the covariance matrix corresponding to the centralized data of each historical time point in each category; determining the first preset number of values in the eigenvectors of each historical time point in each category as the target electrical data of each historical time point in each category; performing prediction analysis on the target electrical data of a plurality of historical time points to obtain target electrical data of a plurality of future time points; performing fault analysis on the target electrical data of each future time point to predict the target fault probability of the electrical equipment at each future time point; determining a maintenance strategy of the electrical equipment at each of the future time instants according to a linear programming method; a loss function of the linear programming method is ; represents a maintenance decision at the th time instant, is 0 represents no maintenance is needed at the th time instant, is 1 represents a maintenance level of 1 at the th time instant, is 2 represents a maintenance level of 2 at the th time instant, represents a specific failure probability at each future time instant determined according to the maintenance decision at each time instant, in a case that the maintenance decision at each time instant all represents no maintenance, the specific failure probability at each future time instant is a target failure probability, represents a maintenance cost corresponding to a target maintenance level indicated by the maintenance decision, represents a failure repair cost;​ performing maintenance management on the electrical equipment according to the maintenance strategy of each future time point.

2. The method of claim 1, wherein, The maintenance management of the electrical equipment according to the maintenance strategy of each future time point comprises: in the case that the maintenance strategy of each future time point includes that maintenance is needed at a target future time point and the maintenance method is to reduce the operating power of the electrical equipment to a target operating power, reducing the operating power of the electrical equipment to the target operating power at the target future time point.

3. The method of claim 1, wherein, The method further comprises: obtaining the specified fault probability of at least one other electrical equipment in addition to the electrical equipment at each future time point; determining the total operating power of the electrical equipment and the at least one other electrical equipment; determining the maintenance strategy of the electrical equipment at each future time point according to the total operating power, the target fault probability of each future time point and the specified fault probability of each future time point.

4. The method according to any one of claims 1 to 3, characterized in that, The method further comprises: performing feature analysis on the original electrical data of a plurality of historical time points to obtain feature data of the plurality of historical time points.

5. The method of claim 4, wherein, The prediction analysis on the target electrical data of a plurality of historical time points to obtain target electrical data of a plurality of future time points comprises: training the parameters in a preset prediction model using the target electrical data of a plurality of historical time points to obtain a trained target prediction model; processing the target electrical data of the i th time point using the target prediction model to obtain the target electrical data of the i+1 th time point; wherein i is an integer greater than or equal to 1, and i+1 is less than or equal to the total number of time points of the plurality of future time points. The target electrical data of the i+1th moment obtained each time is determined as the target electrical data of each future moment.

6. The method according to any one of claims 1 to 3, characterized in that, The prediction analysis on the target electrical data of the plurality of historical moments obtains the target electrical data of the plurality of future moments, including: The random error and the p weight coefficients are obtained; wherein, p is the number of moments of the plurality of historical moments; The p weight coefficients are used to respectively perform weighted fusion on the target electrical data of the tth moment to the t+p-1th moment to obtain t+p target data; t is an integer greater than or equal to 1, and t+p is less than or equal to the total number of moments of the plurality of future moments; The target electrical data of the t+pth moment is determined according to the t+p target data and the random error; The target electrical data of the t+pth moment obtained each time is determined as the target electrical data of each future moment.

7. An apparatus maintenance device characterized by comprising: The device comprises: The prediction analysis module is configured to: combine the original electrical data of the plurality of historical moments and the feature data of the plurality of historical moments to obtain the related data of the plurality of historical moments; initialize three center points, which correspond to the following faults, warnings, and normal states respectively; calculate the distance between the related data of each historical moment and each center point, and assign the related data of each historical moment to the nearest center point; add the related data of the plurality of historical moments corresponding to each center point to obtain a third result; divide the third result by the vector length corresponding to the coordinates of each center point to obtain the coordinates of the new position of each center point; continue to perform the calculation of the distance between the related data of each historical moment and each center point until the center points no longer change or the maximum number of iterations is reached; subtract the average value of the related data of the plurality of historical moments corresponding to each category from the related data of each historical moment in each category to obtain the centralized data of each historical moment in each category; determine the eigenvectors of the covariance matrix corresponding to the centralized data of each historical moment in each category; determine the first preset number of values in the eigenvectors of each historical moment in each category as the target electrical data of each historical moment in each category; and perform prediction analysis on the target electrical data of the plurality of historical moments to obtain the target electrical data of the plurality of future moments. The fault analysis module is configured to perform fault analysis on the target electrical data of each future moment to predict the target fault probability of the electrical equipment at each future moment. The strategy determination module is used to determine the maintenance strategy of the electrical equipment at each future time point based on a linear programming method; the loss function of the linear programming method is... ; Indicates the first Maintenance decisions at any given moment 0 indicates the first No maintenance is required at this time. 1 indicates the first The maintenance level at any given time is 1. 2 indicates the first The maintenance level at that moment was 2. This represents the specific failure probability at each future time, determined based on maintenance decisions at each time point. If maintenance decisions at all times indicate that no maintenance is required, then the specific failure probability at each future time point is the target failure probability. × This indicates the maintenance cost corresponding to the target maintenance level indicated by the maintenance decision instruction. Indicates the cost of troubleshooting; The maintenance management module is configured to perform maintenance management on the electrical equipment according to the maintenance strategy of each future moment.

8. The apparatus of claim 7, wherein, The maintenance management module is further configured to, in a case where the maintenance strategy of each future moment includes that the electrical equipment needs to be maintained at a target future moment and the maintenance manner is to reduce the operating power of the electrical equipment to a target operating power, reduce the operating power of the electrical equipment to the target operating power at the target future moment. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method in any one of claims 1 to 6. The processor executes the computer program to implement the steps of the method in any one of claims 1 to 6.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, which is executed by a processor, implements the steps of the method according to any one of claims 1 to 6.

Images